Reading SMPTE timecode information

XFC

OversamplingApplication Note DK9222-0213-0063

KeywordsSMPTEtimecodeEL3702entertainmentstageshowlighttechnologylight controldecodingtime informationsynchronisingLED

Reading SMPTE timecode information

This application example provides basics on the SMPTE timecode and contains a TwinCAT sample code for

extracting the time information sent by a SMPTE master. Using an EL3702 EtherCAT oversampling terminal

it is possible to “listen” to the analog signal the SMPTE master broadcasts to the slaves. For the sample

code see DK9222-0213-0063_TwinCAT_sample.zip.

PreamblePlease keep in mind that in this document the term frame is used in a double meaning:

– Part or unit of a protocol.

– Snippet or single picture as unit of a movie.

What is SMPTE-Timecode?Defined by the Society of Motion Picture and Television Engineers, the SMPTE timecode is a standard to provide an absolute

time reference for media metadata. Mainly it is used to synchronise different layers in media productions. A SMPTE frame

consists of 80 bits and is always broadcast from a SMPTE master to the slaves. Via an analog signal (±1…10 V), the SMPTE

timecode provides time information in a binary coded decimal identification on an absolute 24-hour time base (hh:mm:ss:ff).

Depending on the signal coding, there are different rates of frames per second (fps); the most common values are 24, 25, and

30 fps.

Broadcast

Audiohh:mm:ss:ff

Videohh:mm:ss:ff

Automationhh:mm:ss:ff

Otherhh:mm:ss:ff

SMPTE masterhh:mm:ss:ff

Fig. 1 SMPTE timecode: Broadcasting an absolute time reference to the slaves

New Automation TechnologyBECKHOFF 1For application notes see disclaimer on the last page

XFC


Protocol details | Duration of a single bitThe first step is to calculate the duration of a single bit in relation to the frame rate. Because of the different picture frame

rates (24, 25, 30), the duration of a single bit differs and is calculated as follows:

Frames per second

30

Duration of one bit Duration of a frame

25

24

4162⁄3 µs

500 µs

520 µs

80 x 4162⁄3 µs = 331⁄3 ms

80 x 500 µs = 40 ms

80 x 520 µs = 412⁄3 ms

Table 1 Duration of one bit

Protocol details | High or low bitThe second step is to count the transitions within this duration. The bits are encoded as biphase: A „0“ bit has a single

transition at the start of the bit period. A „1“ bit has two transitions, at the beginning and middle of the period. So by counting

the transitions within a defined period of time, it can be determined whether the bit is high or low. When the signal is stable

during 417 µs at a rate of 30 fps, it is 0. When there are two constant signals of ~208 µs detected, it is 1.

417 µs 417 µs

“Zero“ “One“

208 µs

transition

Fig. 2 Sample at 30 fps

Protocol details | Frame setupAa a timecode frame always consists of 80 bits, it is 10 byte long. To mark the ending of a frame, the last two bytes follow a

defined sequence called SyncWord: 0011.1111.1111.1101. In table 2 you can see the setup of a complete timecode frame. In

the first row an overview of the time segmentation is given. To enable a fast overview of time-relevant information, the bits

containing other information are marked „x“.


XFC


Bit 0

Bit 1

Bit 2

Frameunits

Frame tens

Secondunits

Second tens

Minuteunits

Minutetens

Hourunits

Hour tens

Bit 3

Bit 4

Bit 5

Bit 6

Bit 7

Sync word(fixed sequence of bits)

Byte 1 Byte 2 Byte 3 Byte 4 Byte 5 Byte 6 Byte 7 Byte 8 Byte 9 Byte 10

1

2

4

8

x

x

x

x

10

20

x

x

x

x

x

x

1

2

4

8

x

x

x

x

10

20

40

x

x

x

x

x

1

2

4

8

x

x

x

x

10

20

40

x

x

x

x

x

1

2

4

8

x

x

x

x

10

20

x

x

x

x

x

x

binary 0

binary 0

binary 1

binary 1

binary 1

binary 1

binary 1

binary 1

binary 1

binary 1

binary 1

binary 1

binary 1

binary 1

binary 0

binary 1

Table 2 Binary coded time information in the SMPTE frame. Bits containing other information than timecode are marked x.

Analysing a frameIn the following figures two sections of a frame are decoded as an example. Figure 3 shows the complete SMPTE frame, figure

4 focusses on a frame detail (the frame units and the frame tens) to analyse the binary coded time information and transfer it

into the decimal based count of frames.

FrameUnits

UserBits

FrameTens

UserBits

SecondUnits

UserBits

SecondTens

UserBits

MinuteUnits

UserBits

MinuteTens

UserBits

HourUnits

UserBits

HourTens

UserBits

SyncWord Frame

UnitsUserBits

FrameTensSyncWord

one SMPTE frame

n+1 n-1n

64 68 72 76 0 4 8 12 16 20 24 28 32 36 40 44 48 52 56 60 64 68 72 76 0 4 8 12

Fig. 3 A single SMPTE sample frame

Bit 0

Bit 1

Bit 2

Bit 3

Frameunits

Frame tens

1

01

2

4

8

1

0

0

1

10

20

x

x

frame units = 9 frame tens = 20

1 0 0 1 0 1 x x

Fig. 4 Bit sequence 1001 | frame unit = 9; bit sequence 01xx | frame tens = 20, as both last bits do not give information about time

To obtain the complete frame value, both values are added: 20 + 9 = 29. It is the 29th frame in this second. Extracting further

time data out of the complete sample frame shown in fig. 3 results in: 05:38:14:29. Having understood the coding of the frame,


XFC


you can now focus on a suitable hardware setup to use in conjunction with the TwinCAT PLC sample code. The setup includes

an external SMPTE timecode master and an EL3702 EtherCAT terminal.

TwinCAT Sample ProgramThe file TwinCAT sample.zip contains a program sample to extract time information out of a SMPTE frame. It was programmed

for the following hardware setup:

– EL3702 | 2-channel analog input terminal ±10 V with oversampling

– EK1100 | EtherCAT Coupler

– CX5010 | Embedded PC series with Intel® Atom™ processor; or any Beckhoff IPC with a higher performance processor,

see chapter Limitations

– Video Binloop | External SMTPE device from Alcorn McBride Inc.

EL37021 | Input1+2 | Input1-3 | GND

SMPTE output9 | SMPTEOUT+8 | SMPTEOUT-5 | GND

CX5010

EK1100

EL3702 SMPTE output(D-sub 9-pin, female)

Video Binloop(Alcorn McBride Inc.)

123

985

Fig. 5 Hardware setup and wiring to connect the timecode master to the EL3702

The SMPTE connector provides both an input and output for the SMPTE interface of the Binloop. The balanced input is 600 ohm

transformer-isolated and will accept levels between 1…5 VPP. The balanced output level defaults to 0 dBV into 150 ohms, but

can be adjusted between 0…4 dBV using the front-panel menu of the Binloop.


XFC


The following table gives an overview about the settings of the Video Binloop, which are not configured in TwinCAT but in the

external device.

SMPTE

Setting Description ValueMode Determines whether the SMPTE Interface will generate timecode, or read external timecode.

Gen

Frame rate The frame rate in which SMPTE is generated. 23,976

Preroll Time Initial time of SMPTE generator when it is firstenabled.

Start Time Time that SMPTE generator returns to after it reaches the end time (when looping)

00:00:00.00

End Time Time when SMPTE generator either stops or loops back to start time.

Loop Mode Determines whether SMPTE generator loops continuously from start to end time.

enabled

Powerup Mode Determines whether SMPTE is enabled or disabled when Binloop powers up.

Start

Restart Mode Determines how SMPTE generator responds when itis told to start again. It can either restart or ignorethe start command.

enabled

Idle Mode Determines if SMPTE output is muted or active when the SMPTE Interface is idle.

SMTPE Active

Determines how long SMPTE reader will free-wheel when external SMPTE is lost.

disabled

Output Level Configures the voltage level (in dB) of the SMPTE output.

4 dBV

00:00:00.00

00:00:00.00

Self Gen Mode

Table 3 Binloop parameters (from manual at Alcorn McBride Inc. website) are not configured in TwinCAT

Note:Theoretically the information is readable at all frame rates, but the sample program is designed for 24 fps and has not

been tested for other frame rates. Furthermore only the time information is read out, not the user bits, the flags and other

information transmitted with the frame.


XFC


TwinCAT sampleIn the file DK9222-0213-0063_TwinCAT_sample.zip, the TwinCAT sample program is given as SMTPE_Timecode.pro.

With the code of the program sample, six steps are provided to extract the time information out of the SMPTE frame:

– Conditioning the EL3702

– Detecting the transitions

– Analysing the transitions

– Identifying high and low bits

– Analysing the bitstream

– Extracting the bits

1 | Conditioning the EL3702The EtherCAT Oversampling Terminal EL3702 listens to the SMPTE output signal and provides an array of values to the TwinCAT

PLC. By default the cycle time of the PLC is set to 1 ms and the oversampling rate of the EL3702 is set to 100. Therefore the

signal is split into samples with a size of 10 µs, and 40…50 samples for each timecode bit can be analysed.

2 | Detecting the transitionsIn the array Tab_EL3702_Doublecycles the two last cycles are stored: the samples of the current cycle (n) and the cycle

before that (n-1). In each PLC cycle, the program edits the current cycle samples (n). The samples of n-1 are used to compare

the current state of the timecode input with the past state. In that way it is possible to check whether a change of sign has

occurred in the periods n-1 to n. When the sign is changing, a transition is detected.

IF Tab_EL3702_Doublecycles[ i-1] * Tab_EL3702_Doublecycles [ i ] < 0

(* The analog signal changes sign --> there is an edge *)

OR Tab_EL3702_Doublecycles [ i-1 ] <> 0 AND Tab_EL3702_Doublecycles [ i ] = 0 THEN

3 | Analysing the transitionsTab_EL3702_DoublecyclesTrig is an array that shows only the transitions of the timecode signal. Only the transitions of the

last cycle samples (n) are managed/handled. Tab_EL3702_Doublecycles contains the samples, generated by the EL3702, of

the last 2 PLC cycles. Only transitions of the past cycle are checked, marked in red in the figure below.


XFC


In the pointer Index_Tab_EL3702, the positions of the transitions within the last cycle are stored. For the next PLC cycle, this

pointer is shifted in order to search for new transitions starting at the position of the last one detected. If there is no further

transition detected, an error occurs in Index_Tab_EL3702_Error.


XFC


4 | Identifying high and low bitsThe duration between two transitions indicates the bit state (high = 1 or low = 0). When transitions (changes in the sign

of the signal’s value) are detected, the time between the transitions is measured and named ConstantTime. The value of

ConstantTime is compared to 330 µs to identify the bit status.

The lower duration limit between two transitions is 417 µs at 30 fps, the upper limit of a half duration is 260 µs at 24 fps (0.5 x

520 µs). This results in a reference of < 330 µs as this fixed limit matches all frame rates when searching for high and low bits:

– If ConstantTime > 330 µs, the transitions include a low bit (0). (“duration long”)

– If ConstantTime < 330 µs, the two transitions measured could be part of a triple identifying a high bit (1). If the next

period is < 330 µs, too, the high bit (1) is identified. (“duration short”)

5 | Analysing the bitstreamTo identify a bit as 1, the Even variable is set to TRUE at the first detection of “duration short”. If Even still is TRUE, as a

“duration long” is detected, an error occurs (FrameError), because the number of consecutive ocurrences of “duration short”

must always be even. Therefore a high bit (1) can only be identified, if a low bit (0) has occurred before. With First_BitZero_

Reach, the detection of a high bit (1) is started only, if a low bit (0) has already been detected before.

The detected bits are stored to the array Tab_TimeCodeFrame which contains the sequence of bits detected in the timecode

signal.

6 | Extracting the bitsAs the timecode frame consists of 80 bits, the next step to extract a valid frame is going through the array Tab_

TimeCodeFrame to find the bit sequence of the SyncWord: If the SyncWord sequence is detected behind position 80 in the

array, a valid frame can be extracted to LastTimeCodeFrame. From here the information can be decoded to TimeCodeFrame

with table 2 to read the information inside.

To prevent overflow of the array TimeCodeFrame, an error is set in Index_Tab_TimeCodeFrame_Error if more than 300

timecode bits are detected without a SyncWord sequence.


XFC


LimitationsThe complete program causes quite a heavy load on a CPU as a lot of data is copied and calculations are executed. Running

the program on a CX5010 at a PLC cycle time of T#1ms occupies 50% of the CPU capacity. When expanding the cycle time

to T#2ms at an oversampling rate of 100, the CPU load will still be 25%. Under these conditions it is however possible to

manage 20…25 samples per timecode bit which is quite acceptable.

Please note: When the PLC cycle time is changed, other constants in the program have to be adjusted, too:

– 10 stands for a 10 µs oversampling rate (cycle time/oversampling rate: 1 ms/1000 = 10 µs). If the cycle time is

expanded to 2 ms, the oversampling rate should be set to 20.

– If the oversampling rate in the source code is changed, the factor has to be adapted in the System Manager, too.

Fig. 6 Further constants needs to be adjusted if the cycle time is changed.

With a PLC at T#1ms and an oversampling rate of 50, the CPU load is still 40%. A further decrease is not possible as then

the number of sample would not be sufficient to generate the signal correctly. Therefore the use of a CX5020 or a higher

performance device is recommended.

Cycle time Oversampling rate1 ms

1 ms

2 ms

50

100

100

CPU load40 %

50 %

25 %

Table 4 Performance limitations with a CX5010


XFC


This publication contains statements about the suitability of our products for certain areas of application. These statements are based on typical features of our products. The examp-les shown in this publication are for demonstration purposes only. The information provided herein should not be regarded as specific operation characteristics. It is incumbent on the customer to check and decide whether a product is suit-able for use in a particular application. We do not give any warranty that the source code which is made available with this publication is complete or accurate. This publication may be changed at any time with-out prior notice. No liability is assumed for errors and/or omissions. Our products are described in detail in our data sheets and documentations. Product-specific warnings and cautions must be observed. For the latest version of our data sheets and documentations please visit our website (www.beckhoff.com).

© Beckhoff Automation GmbH, February 2013The reproduction, distribution and utilisation of this document as well as the communication of its contents to others without express authorisation is prohibited. Offenders will be held liable for the payment of damages. All rights reserved in the event of the grant of a patent, utility model or design.

– ± 10 V analog input terminal with oversampling www.beckhoff.com/EL3702

– Download the TwinCAT sample

http://download.beckhoff.com/download/document/Application_Notes/DK9222-0213-0063_TwinCAT_sample.zip

– see also Application Example for DMX protocol

http://download.beckhoff.com/download/document/Application_Notes/DK9222-0311-0029.pdf

– PC-based Control for Stage and Show Equipment www.beckhoff.com/stage

– PLC and Motion Control on the PC www.beckhoff.com/TwinCAT

– SMPTE time code master ‘Video Binloop’ from Alcorn McBride Inc. www.alcorn.com

New Automation TechnologyBECKHOFF 10

Documents

Reading SMPTE timecode information